Generation of a Real-Life Battery Usage Pattern for Electrical Vehicle Application and Aging Comparison With the WLTC Profile

The process of determination of batteries lifespan for a specific application generally contains some aging tests. Most of the time, these tests used simplified power or current profiles designed specially to be as representative as possible of the reality. The use of a real profile could improve the reliability of the approach but needs to know how to select the typical profile from an available database. This article shows the methodology used to analyze a database of 10 electric vehicles monitored during 2 years. Different mathematical methods (Pearson and Spearman coefficient of correlation, Shannon entropy, principal component analysis, K-means, horizontal ellipsis ) were applied to sort the essential variables and classify the uses of the vehicle. Representative classes of uses were then gathered to constitute a profile used in aging tests. The aging results with this real-life type profile and with an adjusted WLTC (Worldwide Harmonized Light Vehicles Test Cycles) profile are presented and compared. The effectiveness of the use of the WLTC profile for aging tests in real life automotive application is shown.

Automated summary

The article presents a methodology for analyzing data from 10 electric vehicles to select representative profiles for battery aging tests. Mathematical methods were used to sort key variables and classify vehicle uses, resulting in a realistic profile for aging tests. The effectiveness of using a real-life type profile compared to a modified WLTC profile for aging tests in automotive applications is demonstrated.